Introduction to Grey system theory
The Journal of Grey System
Instance-Based Learning Algorithms
Machine Learning
Fundamentals of algorithmics
Wrappers for feature subset selection
Artificial Intelligence - Special issue on relevance
Neural Networks for Pattern Recognition
Neural Networks for Pattern Recognition
Feature Selection for Knowledge Discovery and Data Mining
Feature Selection for Knowledge Discovery and Data Mining
Machine Learning
Toward Integrating Feature Selection Algorithms for Classification and Clustering
IEEE Transactions on Knowledge and Data Engineering
An instance-based learning approach based on grey relational structure
Applied Intelligence
| Hi-index | 0.00 |
In this paper, a novel hybrid Taguchi-Grey-based method for feature subset selection is proposed. The two-level orthogonal array is employed in the proposed method to provide a well-organized and balanced comparison of two levels of each feature (i.e., the feature is selected for pattern classification or not) and interactions among all features in a specific classification problem. That is, this two-dimensional matrix is mainly used to reduce the feature subset evaluation efforts prior to the classification procedure. Accordingly, the grey-based nearest neighbor rule and the signal-to-noise ratio (SNR) are used to evaluate and optimize the features of the specific classification problem. In this manner, important and relevant features can be identified for pattern classification. Experiments performed on different application domains are reported to demonstrate the performance of the proposed hybrid Taguchi-Grey-based method. It can be easily seen that the proposed method yields superior performance and is helpful for improving the classification accuracy in pattern classification.